VARIABLE SPEED REGULATION FOR PUMP AS TURBINE IN A …

IAHR 2019E-proceedings of the 38th IAHR World Congress September 1-6, 2019, Panama City, Panama
doi:10.3850/38WC092019-1200
4039
VARIABLE SPEED REGULATION FOR PUMP AS TURBINE IN A MICRO PUMPED HYDRO ENERGY STORAGE APPLICATION
ALESSANDRO MORABITO(1), GILTON C. A. FURTADO(2),
ANDRE L. AMARANTE MESQUITA(3) & PATRICK HENDRICK(1)
(1) Université libre de Bruxelles (ULB), Aero-Thermo-Mechanics Dept., École Polytechnique, Brussels, Belgium
[email protected]; [email protected] (2) Eletronorte, State of Pará, Brazil [email protected]
(3) Núcleo de Desenvolvimento Amazônico em Engenharia, Universidade Federal do Pará, State of Pará, Brazil
[email protected]
ABSTRACT
The need of storage technologies is recently emerging and more discussed in micro scale. Pumped hydro energy storage (PHES), the most used technology in energy storage, represents a valid option for its maturity, proven long life-span and rather high efficiency. Using a centrifugal or diagonal Pump As Turbine (PAT) in the hydroelectric generation is a valid trade-off between capital cost and performance. However, commercial pumps are not designed to run in reverse mode. Hydraulic efficiency and working range are so limited. Variable speed regulation allows the PAT to maintain a nearly constant high efficiency all the time regardless of the head availability or the instant surplus of energy and enlarge the operability range in pumping and generating modes. This paper presents an overview of an existing micro-PHES installation of about 17 kWh capacity. This facility is equipped with a PAT coupled with a Variable Frequency Driver (VFD), which allows the system to deal with variable load keeping high efficiency. The characteristic curves of the pump in reverse mode are analysed defining the exploitable PAT working range. The use of variable speed pumping in PHES integrated into a smart grid aims to give flexibility in storing energy. The extraordinary adaptability recorded on the site in partial load until 30% clearly endorses the coupling of the PAT and the variable speed driver.
Keywords: Pumped Hydro Energy Storage (PHES); Variable Speed; Hydraulic System Performance; Smart Grid; Pump As Turbine (PAT).
1 INTRODUCTION In 2016 renewable energy sources accounted more than half of the new power capacity installed worldwide,
with almost 165 gigawatts coming online (IEA, 2017). This has been driven by a sharp cost reduction on the renewable technology sources and by policy support around the world, especially for photovoltaic (PV) capacity which grew by 50% last year, reaching over 74 GW (IEA, 2017). Moreover, for the first time, the solar PV additions are growing faster than any other fuel, surpassing the net growth in coal (IEA, 2017; Mohd, 2008). In an accelerated case, where government policy disrupts barriers to growth, IEA analysis finds that renewable capacity growth could be boosted by another 30%.
However, an important drawback emerges: solar and wind energy have to deal with production intermittency due to the availability and thus on seasonally or daily weather. A frequent mismatch between renewable energy production and instant use then might occur. On the other hands, in remote area, this usually leads to install fossil-fuel energy sources, not environmentally helpful, to supply base-load demand, e.g. diesel engines.
Besides the large-scale power system, micro-energy systems become promising alternative towards energy reliance of the people as in rural environment as in not isolated area. Representing such a relevant share of energy consumption, urban buildings play an important role in saving energy and in the dependence of fossil fuel. The recent adopted policies in Europe and worldwide, enforcing the decarbonisation of energy supply systems, have highlighted the pivotal role of flexible loads on balancing the grid (IEA, 2017). A conglomerate of utilities, named Smart Grid, aims to keep a constant interaction in the form of energy, mass and information between the building and its surroundings dealing with the growing adoption of decentralized renewable energy sources. Energy storage is one of the different operations and approaches that aim at making Smart Grid flexible and controllable. Among energy storage technologies currently available, Pumped Hydro Energy Storage (PHES) is recognized as one of the most cost-effective because of its predictable energy characteristics, its long-term reliability and its reduced global environmental effects (Steimes, 2016). During off-peak demand, PHES pumps back water to an upper reservoir from a lower reservoir and stores potential energy exploitable
E-proceedings of the 38th IAHR World Congress September 1-6, 2019, Panama City, Panama
by hydraulic turbines when it is needed. The performance of the site is unequivocally dependent on the selection of the turbomachinery adopted and it is case sensitive: pumps and turbines have to be installed according to exploitable flow rate, available head and pipeline system characteristic curve. Although pumps and turbines are designed for a particular flow rate, they operate across a certain range due to fluctuation of the available head and/or load. The water level might variate along the full cycle of emptying and filling the reservoir according its specific geometry and the flow regime.
The micro energy grid, presented in this paper and illustrated schematically in Figure 1, has different renewable energy sources in order to support the energy consumption of heating, ventilation, and air conditioning of the two connected buildings, mostly dedicated to offices and conference rooms. A solar flower of 5.2 kW peak and other groups of photovoltaic panels are installed (Morabito, 2019). Four wind turbines of 2.4 kW each are also present. In order to store energy from the renewable sources a micro PHES, equipped with a single PAT (Pump As turbine), is installed in the smart grid. The capacity of the micro system is about 17 kWh and it reaches power production peak of 7 kW. The maximum energy store capacity is increased by the implementation of a vanadium redox-flow battery of 100 kWh at 10 kW in the smart grid.
To design PHES, the capacity of the site has to be defined according to the available spaces and geodetic head. On site, the lower reservoir limits the maximum capacity at about 650 m3 of water. Due to its geometry (deep and narrow) it may occur to register a large variation of the water level in the reservoir during the normal working period. The available head usable by the pump in reversed could drop down to less than 50 % during the generating phase. For this reason, the pump is coupled with a Variable Frequency Driver (VFD), which allows the system to deal with variable load keeping high efficiency. The use of variable speed pumping in PHES integrated into a smart grid aims to give flexibility in storing energy. According to the instant green energy surplus provided the renewable power sources, the pump rotational speed could be set to raise the energy level independency of the smart grid.
The objective of this paper is to explore the application of variable speed in this technology, showing its advantages in increasing considerably the hydraulic efficiency and its flexibility in variable load. Relevant design considerations are discussed in terms of PAT selection methodology, hydraulic system performance and on the use of variable speed.
Figure 1. Schematic representation of the micro PHES case study in the smart-grid.
2 PUMP AS TURBINE Pumps behave differently when they work in reverse in terms of forces, flow and power (Williams, 1994). A
pump increases the fluid pressure by the conversion of mechanical energy; in the opposite rotational direction, the PAT exploits the head available to generate power. Since the rotational speed and the flow direction are inverted in turbine mode, the velocity triangles change as a consequence. Small and commercial pumps do not have particular reason to be designed in reversed mode unlike pump-turbines used in large PHES. Pumps are not usually equipped with guide vane usable in generating mode for directing the flow. In PATs, the casing works as the water flow guide. Figure 2 illustrates qualitatively the complete characteristics of a pump in four quadrants. The normal operation of a pump is represented with a positive rotational speed, N+, and positive flow rate, Q+. Only one sector of this sector corresponds to actual pumping. The darker areas in the diagram of Figure 2 represent the operating conditions where the pump dissipate energy and only thermal energy is generated. Increasing the flow rate from a normal pump operation, eventually the pump is not able to deliver a positive head. H < 0 expresses the higher
E-proceedings of the 38th IAHR World Congress September 1-6, 2019, Panama City,
Panama
4041
pressure at the pump inlet than at the pump nozzle: crossed dissipation area, the pump runs as an abnormal turbine, namely reverse flow turbine. Here the pump runs still at N+ and Q+ but with opposite gradient of pressure. However, PAT operation is located at H > 0 condition with negative rotational speed, N-, and negative flow rate, Q-. Under H > 0 (higher pressure at the pump outlet than at the inlet), PAT operating conditions can range according to the characteristics defined by the rotational speed NT1, NT2, etc. The effective characteristic curves are limited by the runaway curve for torque, M, equal to zero. A second operational limit is reached by the characteristics of the hydraulic resistance at N = 0 (Figure.2b). This curve is the result of water friction inside a no-rotating runner.
Figure 2. Complete pump characteristics in four quadrants defined by the rotational speed, N, and the flow rate, Q. Normal pump (2.a) and PAT characteristics (2.b) showed in H-Q plot
Thoma and Kittredge (1931) first investigated the use of PATs in the last century in order to recover energy from high-pressure circuits in heavy industry or chemical systems. The common alternative before that was to throttle the flow and waste energy. For decades researchers discussed the hydraulic behavior of pumps in reverse mode and tried to formulate a function or a model able to define the hydraulic performance. A collection of proposed correlations is described in numerous reviews present in the literature (Binama, 2017). When the pumping system is no longer able to efficiently deliver a positive flow rate against a static head, the minimum operating pump shaft speed is defined in respect of the practises for a good pump reliability at each speed (Figure.3). Design engineers define the constraints for the system, such as flowrate, NPSH margin, or fluid velocity to preserve the reliability of the system (Barringer, 1997). The characteristic defined by Mean Time
Between Failures (MTBF) aside the chart in Figure 3 estimates the probability that the pump will operate
for its expected lifetime without a failure. The extreme operating conditions affect the total cost of the
system and should be avoided. Hence, the minimum operating speed to each working condition is specific. In reverse mode, according to the velocity flow angles at the inlet and outlet of the runner at the selected rotational speed, PATs are able to deliver a power-output or a positive torque (M > 0) only above a minimum flow-rate QMIN as also showed in Figure 2.b. Below this value, the PAT power output is negative. In other words, for QMIN and consequently at H < HMIN, the power station is actually dissipating energy to maintain the runner at the constant NT. If possible, N should be re-set, when appropriate, updating the minimum values of QMIN and HMIN.
3 SYSTEM CONFIGURATION The performance of the pump in the hydraulic system is determined by pipeline characteristics and size. In
the same way, the maximum energy that a turbine can extract from a fall of water is on the net of the pressure losses generated by friction. The energy of water flowing between two points is expressed by Bernoulli equation. Pressure losses appear along the pipeline (linear pressure drop) and in valves, bends and other piping fittings as for obstructions, which are called local pressure losses.
In this case study about 80m underground polyethylene pipe (355 mm of external diameter) connects the upper reservoir to the technical local where the fittings are in steel and with smaller diameter. The two pipe inlets
4042
to the reservoirs are equipped with a robust grid to protect the impeller from large dangerous objects. A periodical screening of their condition is scheduled to ensure clean passages. Two valves are needed to isolate the system and allow maintenance. A third valve is installed in the technical room for managing the operation of the site.
Figure 3. Reliability impact (MTBF) of operation away from BEP to ANSI pump. Adapted from Barringer, 1997.
According to the Colebrook-Prandtl equation, the head loss due to the losses in straight pipe (in polyethylene) is about half meter at the maximum measured flow rate. The local pressure losses are mostly produced by the fittings to the pipeline of each reservoir. In order to reduce these undesirable effects two pipes with 6 degrees of convergence are installed on each side of the pipeline. Two grids protect the pump from external object at the price of further energy losses. The exploitable pressure gauge left is depicted by the energy grade line which is a plot of the energy head over the horizontal distance. Figure 4 describes the energy grade profiles during the hydropower generation by mean the PAT. The pressure sensors are installed as follow: at the discharge side of the PAT (sensor D), before the PAT (sensor U) and before the regulating valve at upper-reservoir side (sensor C). Sensors D and U are dedicated to measure the total pressure gap from downstream and upstream the PAT. The sensor C is able to control and indicate the exploitable water level from the upper reservoir while the main valve is close. The available head reduction due to the distributed losses in the straight pipe (in polyethylene) is about 0.6 m at the maximum measured flow rate. Local pressure losses are also produced by the different fittings of the pipeline. Two grids protect the pump from external objects at the price of further losses (about 0.2 meters each) and a flanged diffuser is necessary to connect two pipes of different diameters. In order to reduce these undesirable effects two diffusers are installed on each pipeline inlet.
Panama
4043
Figure 5. Graphical representation of the energy grade line of the system in PAT mode and schematic positioning of the pressure sensors C, U and D.
4 VARIABLE SPEED The pumping station is capable to use only the surplus of energy given intermittently and in different amount
by the rest of the system. Indeed, the pump is capable to use a wide range of electrical power moving from 5 kW (low speed) up to 17 kW (high speed) for storing energy in the upper reservoir. As speed changes, efficiency follows parabolic curves with their apexes at the origin of the graph. Thus, except for a slight shift dependent by Reynold numbers, efficiency is independent by rotational speed. Referring to the nominal speed of the installed pump, the minimum operating pump speed is at 65% when the lower reservoir is full (minimum static head) and at 90% when the lower reservoir is about to be empty (minimum static head). Figure 4 pictures the hydraulic efficiency gain by using the pump with variable rotational speed compared to fix nominal rotational speed (1000 rpm):
= ηVFD − η1000
In this paper the head is normalised based on the nominal pumping head. Rotational speed is normalised on the nominal value in pump mode: N11 = N/N0. NPUMP needs to accordingly adapt to the system required head. The pump working range stays between N11 = [ 0.960, 1.16]. This method allows the installed pump to cover required head not accessible by using the nominal rotational speed N0 and, thus, to record very high efficiency gains.
Figure 4. Gain in hydraulic efficiency in pump mode by using variable speed over the fluctuation of the head.
4044
The primary goal of using the installed PAT with variable speed is to maintain a nearly constant high efficiency all the time regardless of the head availability. Practically the two water levels are always variating because the limited space in the reservoirs. At the minimum load conditions, the PAT runs with a static head reduction of about -55% from its maximum. Lower rotational speed allows the PAT to adapt to the running water regime at reduced available head and to preserve high efficiency in partial load. As shown by Figure 5, the use of PAT at NPUMP = N0 (black line) is clearly not sustainable nor rentable. At this working condition, the PAT could reach only the 73% of the efficiency peak measured in a very limited operation range. However, the PAT records better performance at the selected speed of N11, PAT = 0.6 (blue line): it reaches its maximum for normalised head equal to 0.7 and preserve high values for greater available heads. High vibrations are detected at lower head and they suggest a consequent adjustment of the rotational speed. If it were not for the working condition variations, a gear-box would provide two rotational speeds on the shaft: one suitable for pumping and the second applied for generating mode, but any speed adjustment would be possible. The green line in Figure 5 depicts the efficiency line of variable speed command though a wide range of available head. The difference in efficiency in turbine mode of operating at the variable speed and fixed speed of N11, PAT = 0.6 is defined
= ηVFD − η0.6
Figure 6 describes the gain in hydraulic efficiency by using variable speed over the fluctuation of the available head, showing the rotational speed path to obtain the maximum efficiency in whole operating range. The gain in production and optimized life cycle cost (i.e. avoiding dangerous working conditions) would balance the cost of the VFD and allow a longer life expectancy.
Figure 5. PAT efficiency normalized to its maximum over the available head for different speed regime.
Figure 6. Gain in hydraulic efficiency by using variable speed over the fluctuation of the available head.
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalised Available Head [-]
PAT at constant N = N11 0
Panama
4045
The tests of the pump in reverse mode conducted in micro PHES lead to very favourable outcomes for these applications. The maximum efficiency measured in reversed mode is approximately 2% less than the efficiency of the turbo-machine in pump mode. For this reason, PAT may represent a valid option to the most common hydraulic turbines used for micro scales.
For larger installations, PATs of different sizes can be used in parallel to provide modulated power output: the smaller will define the resolution of the modulated output while the sum of all the machines is the maximum value. Another alternative is coupling to a system at constant N, dedicated to base the load, with only one PAT at variable speed to modulate the total output: this will avoid equipping each pump with a driver.
Variable speed control adds extra high cost and complexity. The cost of the driver can approximately equal the pump cost (Morabito, 2019). Moreover, the efficiency of the driver must be also considered when evaluating the net power produced. On the other hand, the major advantages of variable speed generation, besides the objective of succeeding in load in fluctuation, are:
Fewer machines are needed with saving in space for machine room and less equipment to maintain.
The high inrush current at the start is remarkably reduced.
Gradual changes are produced in the flow that do not upset the normal functioning of the system: pump and PAT can be ramp-up (or down) so that water hammer is reduced.
The soft and reduced number of the starts do not cause power dips and help the system maintenance, cutting the wearing on bearings and flexing of the shaft.
4 CONCLUSIONS Decentralized electricity storage provides a mechanism to control the consumption linked to the total
electricity profile. The main focus should be on micro-small storage systems implemented near to the decentralized energy sources which are quickly growing nowadays. The example of flexibility provided by the case study depicted in this paper can constitute an alternative to upgrading the distribution grid: the load outweighs are cut off by exploiting the capacity of the PHES and the intermittent energy injections produced by the renewable sources are properly stored in the hydraulic reservoir. This is successfully obtained by installing a PAT with variable speed control: it provides conditions for achieving a higher energy yield while avoiding the onset of operation instabilities caused by the part load or by the PAT full load operation.
PAT at variable speed is the valid turbomachinery alternative to conventional micro turbines in exploiting variable hydraulic power. The use of the same machine for pumping and generating is related to the objective of saving cost of energy, space and maintenance. The PAT hydraulic efficiency recorded in the case study (about ~72 %) may be slightly smaller than with regular micro turbines but still very competitive.
In this paper emerges that PHES requires a calibrated evaluation of the hydraulic system performance taking in account energy fluctuation and the turbo-machinery working condition. The extra cost applied for VFD is mitigated by the advantage of installing a PAT instead of an expensive customized micro hydraulic turbine. The extraordinary adaptability recorded on the site in partial load until 30% clearly endorses the coupling of the PAT and the VFD.
ACKNOWLEDGEMENTS This research was supported by the Walloon Region in the framework of the SmartWater project. Authors wish to acknowledge IDETA and the DG04-DG06 for their continuous support.
REFERENCES Barringer H. P., (1997). Reliability Engineering Principles. Barringer & Associates, Humble, Texas. Binama, M., Su, W. T., Li, X. B., Li, F. C., Wei, X. Z., & An, S. (2017). Investigation on pump as turbine (PAT)
technical aspects for micro hydropower schemes: A state-of-the-art review. Renewable and Sustainable Energy Reviews, 79, 148-179.
IEA, (2017). Market and Report Series: Renewables 2017, OECD/IEA, International Energy Agency, Paris. Mohd, A., Ortjohann, E., Schmelter, A., Hamsic, N., & Morton, D. (2008). Challenges in integrating distributed
energy storage systems into future smart grid, Industrial Electronics, IEEE International Symposium. Morabito, A., & Hendrick, P. (2019). Pump as turbine applied to micro energy storage and smart water grids: A
case study. Applied Energy, 241, 567-579. Steimes, J. & Hendrick, P. (2016). Cost and revenue breakdown for a pumped hydroelectric energy storage
installation in Belgium, Sustainable Hydraulics in the Era of Global Change 269–276. Thoma, D. & Kittredge, C.P. (1931). Centrifugal pumps operated under abnormal conditions, Power 73, 881-
884 Williams, A. A. (1994). The turbine performance of centrifugal pumps: a comparison of prediction
methods. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 208(1), 59-66.
4046
Figure 4. Gain in hydraulic efficiency in pump mode by using variable speed over the fluctuation of the head.
Figure 5. PAT efficiency normalized to its maximum over the available head for different speed regime.

VARIABLE SPEED REGULATION FOR PUMP AS TURBINE IN A …

Documents

Transcript of VARIABLE SPEED REGULATION FOR PUMP AS TURBINE IN A …